Early Prediction of Heart Failure From Routine Cardiac CT Using Radiomic Phenotyping of Epicardial Fat.

Area of Science:

• Cardiology
• Radiology
• Artificial Intelligence
• Medical Imaging

Background:

• Epicardial adipose tissue (EAT) is a metabolically active fat depot influencing myocardial biology.
• EAT composition changes in response to cardiac signals, indicating potential for risk stratification.
• Coronary computed tomographic angiography (CCTA) offers an opportunity for noninvasive EAT assessment.

Purpose of the Study:

• To develop and externally validate a reproducible radiomic signature of EAT associated with incident heart failure (HF).
• To investigate the potential of radiomic characterization of EAT for early HF risk stratification.

Main Methods:

• A multicenter cohort study involving 72,751 adults without known HF undergoing CCTA.
• Automated segmentation of EAT and extraction of 1,655 radiomic features.
• Development of a fat radiomic profile for HF (FRP_HF) using a survival autoencoder, with external validation in a separate cohort.

Main Results:

• The FRP_HF demonstrated robust discrimination for incident HF in both internal (C-statistic: 0.869) and external (C-statistic: 0.850) validation cohorts.
• Each 25-percentile increase in FRP_HF was associated with a nearly 4-fold higher adjusted HF risk.
• In the external cohort, FRP_HF significantly improved the discrimination and reclassification of 5-year HF risk beyond conventional models.

Conclusions:

• Automated radiomic phenotyping of EAT from routine CCTA enables scalable, biologically informed stratification of future HF risk.
• Opportunistic imaging-based visceral fat profiling using FRP_HF shows potential as a tool for precision HF prevention.
• This approach allows for early identification of individuals at high risk for HF before clinical manifestation.